| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Developments |
1 From the Depts of Radiology (N.H., M.J., D.K., A.N., S.G.S., R.K., F.A.J., C.M.C.T.) and Radiation Oncology (R.C., A.V.D.), Brigham and Women’s Hosp and Harvard Medical School, 75 Francis St, Boston, MA 02115; and AI Laboratory, Massachusetts Inst of Technology, Cambridge (D.G.). Received Mar 14, 2000; revision requested May 2; final revision received Jan 10, 2001; accepted Feb 6. N.H., R.K. supported by NSF 9731748 Engineering Research Center. R.K., F.A.J. supported by NIH/NCI 1P41RR13218-01 Neuroimaging Analysis Center and NIH/NCI 5P01CA67165-03 MR Guided Therapy. Address correspondence to C.M.C.T. (e-mail: ctempany@bwh.harvard.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and MethodsResults DiscussionREFERENCES |
|---|
Index terms: Magnetic resonance (MR), guidance • Magnetic resonance (MR), image processing • Prostate, biopsy, 844.1261 • Prostate, MR, 844.121411, 844.121412 • Prostate neoplasms, MR, 844.121411, 844.121412, 844.32
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults DiscussionREFERENCES |
|---|
In men with repeatedly negative transrectal US-guided biopsy results and increasing PSA levels, the possibility of an occult cancer always exists. For these men, and for the smaller groups that cannot undergo transrectal US-guided biopsy due to prior rectal surgery, there is a need for an alternative biopsy approach. One such approach is to use computed tomography (CT) (10). However, since we have performed magnetic resonance (MR) imaging–guided percutaneous biopsy at other sites (11), we decided to evaluate MR imaging-guided prostate biopsy.
MR imaging can clearly depict not only the prostate itself but also its substructure including the peripheral zone and, on T2-weighted images, can demonstrate nodules in the peripheral zone, although, the specificity for diagnosis of cancer is limited. Since the peripheral zone is the most common origin site of prostate cancer among the three prostate zones (peripheral, central, and transitional), localizing and targeting the peripheral zone and tumor foci in prostate biopsy may increase cancer detection rate.
Localizing and targeting tumor foci and the peripheral zone with MR imaging before or during prostate biopsy may increase the cancer detection rate (12). Recently, Perotti et al (13) used endorectal MR imaging findings of suspected tumor foci to guide the placement of needles during transrectal US-guided biopsy. By localizing suspected tumor lesions or targets on the endorectal MR image and by visually correlating the locations to US images during transrectal US-guided biopsy, they found that the accuracy of the transrectal US-guided biopsy, aided by using MR imaging, was 67% in a study of 33 patients.
We propose a method to use MR imaging for prostate biopsy, not only to localize tumors and the peripheral zone, but also to place the needles into focal lesions with direct MR imaging guidance. The purpose of our study was to validate and report the feasibility of MR imaging–guided prostate biopsy using surgical navigation software. This biopsy of the tumor foci and peripheral zone is possible by adapting the technical capabilities of MR imaging–guided prostate brachytherapy in an open-configuration MR imager (14) and by implementing surgical navigation software originally developed for neurosurgery (15,16).
|
Materials and Methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults DiscussionREFERENCES |
|---|
In both men, the MR imaging examination before biopsy showed a suspected tumor in the prostate. Both patients had prior histories of proctocolectomies for ulcerative colitis. The first patient (case 1) was a 59-year-old with a PSA level of 6.1 ng/mL on August 6, 1998 and 8.8 ng/mL on March 11, 1999. The second patient (case 2), a 76-year-old, had a increasing PSA level from 4.1 ng/mL in 1989 to 43.5 ng/mL on August 19, 1999. The nature of the biopsy procedure was discussed with the patients, and informed consent was obtained. The study was performed as part of our MR imaging–guided interventional program for which we have institutional review board approval.
Validation Studies
We performed phantom validation studies to assess the tissue discrimination capability of a single T2-weighted fast spin-echo (FSE) section resampled from previously obtained volumetric T2-weighted images and to objectively measure the accuracy of matching between real-time fast gradient-recalled echo (FGRE) and resampled T2-weighted FSE images in the second part of the study. We used a quality assurance phantom (GE Medical Systems, Milwaukee, Wis) that was a cylindric plastic container filled with a solution of doped water with characteristic subcontainer markers (lines, shapes, letters) inside.
The first part of the validation studies involved two sets of data: (a) five real-time FGRE images (transverse, 24.5/12.1 [repetition time msec/echo time msec], field of view of 24 cm, section sickness of 5.0 mm, matrix size of 256 x 128, one signal acquired) at randomly selected locations of the phantom; and (b) their corresponding, real-time T2-weighted FSE images re-sampled from volumetric T2-weighted FSE images (4,050/135; field of view of 24 cm; section thickness of 4 mm; section gap of 0 mm; 256 x 224 matrix; three signals acquired). Intensity profiles of the two randomly selected lines cutting through the markers were generated from both FGRE and T2-weighted images to compare the location of intensity decreases and/or increases at the boundaries of the markers.
In the second part of the study, we created a composite of real-time FGRE and T2-weighted images to investigate the possible discrepancies between the two images, which could be due to the reconstruction inaccuracy or chemical shift. Five readers as a group measured the degree of image matching by classifying the continuity of the marker boundary on consecutive cell borders according to no gap, 0.5-pixel gap, and 1-pixel gap performed at all the crossing points of the cell borders and the marker boundaries.
Preoperative 1.5-T MR Imaging Study
Each patient underwent an MR imaging examination by using a pelvic phased-array coil and 1.5-T MR imaging unit (Signa LX; GE Medical Systems). The images were analyzed to evaluate the size, location, and substructure of the gland and to identify any suspected lesions. The 1.5-T T2-weighted images were FSE images (4,050/135; field of view of 14 cm; section thickness of 4 mm; section gap of 0 mm; 256 x 224 matrix; three signals acquired). These images provided planning information for the biopsy, both for target definition and depiction of the prostatic zonal anatomy.
Intraoperative Imaging Equipment
The biopsy procedures were performed in the open-configuration 0.5-T MR imaging unit (Signa SP; GE Medical Systems), referred to as an intraoperative MR imager (17). The unit provides real-time MR imaging in the bore while allowing the physicians access to the interventional field from a gap between two vertical superconducting coils. An imaging workstation (Ultra 30; Sun Microsystems, Mountain View, Calif) for the surgical simulation and navigation software (3D SLICER; Massachusetts Institute of Technology Artificial Intelligence Lab and Brigham and Women’s Hospital, Boston, Mass) was set up next to the imager console in our interventional MR imaging suite (15). The workstation is connected to two monitors driven by separate graphics accelerator cards (Creator 3D; Sun Microsystems), both of which are installed inside the bore of the magnet or in-bore. One video output is intended for the computer operator display, and the other is for the physician performing the procedure. This display signal is converted to an North American Television Standards Commission video signal and passed to an in-bore monitor placed next to the other in-bore monitor displaying images from the MR imager.
3D SLICER (Fig 1) is surgical simulation and navigation software that displays multimodality images two and three dimensionally. 3D SLICER used with MR imaging–guided prostate biopsy has the ability to simulate real-time T2-weighted images, which are otherwise not practical with the intraoperative MR imager. The simulation was accomplished by resampling the preloaded 0.5-T T2-weighted FSE images obtained a few minutes before the needle placement. The resampling plane of the simulated real-time T2-weighted image should be at the same imaging location and direction of the single section real-time FGRE images.
|
The imager and the imaging workstation have a transmission control-to-internet protocol interface for online and real-time image transfer. The speed is 10 Megabits per second for each machine on the switched Ethernet and 155 Megabits per second through the Asynchronous Transfer Mode equipment; thus, each workstation communicates maximally at 10 Megabits per second to another workstation.
Research personnel retained to perform and/or assist in the biopsy included one or two radiologists (C.M.C.T.), an anesthesiologist, a technologist operating the imager, a surgical and a nonsurgical nurse, and a computer operator (N.H.) controlling the 3D SLICER.
Patient Preparation
We elected to perform biopsy with general anesthesia, since targeting specific foci requires that there be only minimal motion of the pelvis. Several days prior to the biopsy, the patients were examined by the anesthesia service, which is standard practice for all men undergoing anesthesia in an MR environment (18). On the day of the procedure, patients came to the intraoperative MR imaging suite and were prepared for the biopsy. Patients were positioned in the lithotomy position, with the table placed in the vertical gap of the imager.
A needle guidance template was set against the perineum and attached to the table. The template, originally developed for prostate cancer brachytherapy, has a grid of 0.0059-inch holes spaced 5-mm apart. The template was registered to the intraoperative MR imager by using the optical tracking system (Flashpoint 5000; Image Guided Technology, Boulder, Colo) integrated to the imager (11). This registration achieves geometric correlation between the template and patient anatomy, enabling the selection of a hole that will guide a biopsy needle toward the target determined from the images. The usage of the template, including the detail of the registration, is presented in reference 19.
We performed T2-weighted FSE imaging (transverse and coronal; 6,400/100; field of view of 24 cm; section thickness of 3.5 mm; section gap of 0 mm; matrix of 256 x 128; two signals acquired) in the intraoperative MR imager with a flexible external pelvic wraparound coil. These 0.5-T T2-weighted images were correlated with the 1.5-T T2-weighted images to define the peripheral zone; any suspicious lesions were identified and localized on 0.5-T T2-weighted images. We measured the target coordinates so we could stereotactically approach the target by using the template. The 0.5-T T2-weighted images (transverse and coronal) were loaded into the 3D SLICER to provide guidance into the peripheral zone and suspected tumor lesions during needle insertion.
Biopsy Method
We placed the needles by using the two guidance methods: first, using real-time FGRE imaging (24.5/12.1, field of view of 24 cm, section thickness of 3.5 mm, 256 x 128 matrix size, one signal acquired) and second, using T2-weighted images processed with the 3D SLICER.
Real-time FGRE images (single section image updated every 8 seconds) were used to guide the needle and to confirm that the target was in the path of the needle. The T2-weighted images were also updated every 8 seconds and were used to find the T2-weighted information elucidating the position of the peripheral zone and the suspected lesion(s).
The sextant and targeted biopsy of suspected lesions were performed with a MR imaging–compatible single action biopsy needle (Single Action Biopsy Device; U.S. Biopsy, Franklin, Ind). The 18-gauge needle had a 20 mm throw and trocar tip. For each needle placement, a hole that will guide the biopsy needle toward the target was selected, and the physician was notified about target depth. Each specimen was carefully labeled based on the location it came from prior to submission to the department of pathology.
The pathologic results for all biopsy specimens were recorded, and full analysis included documentation of the pathologic presence of tumor. If a cancer was diagnosed the Gleason score of the specimen was recorded.
Procedure Time
The entire procedure was 2–3 hours. The time approximated for each step was as follows: 15–30 minutes for patient setup and positioning, 10 minutes for initial imaging, 30 minutes for induction of anesthesia, 5 minutes for T2-weighted imaging, 10 minutes for planning trajectory, 30 minutes for needle placement, and 20 or more for recovery.
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults DiscussionREFERENCES |
|---|
|
|
|
|
The availability of real-time T2-weighted images with 3D SLICER was also valuable in case 2 (Fig 5), where the real-time T2-weighted images clearly identified the same suspected foci as the 1.5-T images, which was not visible on the real-time FGRE images. Six sextant targets and two additional targets (0.1 mm in diameter and 0.5–1.6 cm in length) from suspected lesions in the left apex and left midgland were collected. Tissues from the suspected targets in the left apex and left midgland in this patient showed prostatic adenocarcinoma (Gleason score, 3+3) involving two of three cores(50% of prostatic tissue). The pathologic finding from the sextant biopsy showed no tumor.
|
|
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults DiscussionREFERENCES |
|---|
MR imaging–guided intervention, however, as an evolving technique, is not intended to prevail over transrectal US-guided biopsy at this rudimentary stage in its application. Indeed, it is not the intention of its developers to replace the more simple, effective, reliable, and economic transrectal US-guided biopsy with MR imaging. Rather, this method is intended to provide a unique alternative to transrectal US-guided biopsy in the patients with prior negative transrectal US biopsy finding and a increasing PSA level, and in the smaller group who have undergone abdominoperineal resection.
Because the procedure takes a relatively long time, we ultimately hope to perform it with local anesthesia. We are also aware that the costs of our approach are higher than transrectal US–guided biopsy primarily due to prebiopsy examination with 1.5-T MR imaging to maximize our ability to target a suspected lesion. It is beyond the scope of this work to assess the cost of MR imaging intervention or to analyze the outcomes of the studies being conducted. However, it should be recognized that in men with prior negative biopsy findings this method might, in fact, be a cost-saving option, since repeated transrectal US biopsy over a long period might be more costly. It may also be a quicker way to reverse a diagnosis in a group of men strongly suspected of having prostate cancer. Further investigations on outcome and cost are necessary to validate the value of the method in the larger patient population.
We have previously performed transgluteal percutaneous MR imaging– and CT-guided biopsy of the prostate gland (10). While this approach is feasible, it is limited by the long trajectory and distance from the skin to the gland. This long distance precludes accurate navigation and control of the needle. On the basis of our brachytherapy and biopsy experience, we believe the biopsy should be directed through the perineum (14). The access route is direct, and the sampling of the peripheral zone can be maximized. By using this approach, the volume of the peripheral zone in each sextant biopsy will likely be increased, if compared with transrectal samples. While we have not yet evaluated this feature, we believe this will be an added benefit to the ability of MR imaging to target specific lesions in the gland.
Although our preliminary results support MR imaging–guided biopsy as a promising intervention, the success of the procedure relies on its ability to localize suspected lesions and to depict the peripheral zone during 0.5-T intraoperative MR imaging. A previous study (20) examined this issue by comparing the image quality in a subset of 20 patients, evaluating each of the images for ease of identification of the gland, its substructure, and abnormal foci with the endorectal coil 1.5-T images and the pelvic wrap 0.5-T images. In almost all aspects, the 0.5-T images compared favorably with the 1.5-T images. The sole difference was found in the identification of focal lesions; that is, in 12 of 20 men, foci were identified at 1.5-T and not at 0.5-T alone. For this reason, we need to refer to the preoperative 1.5-T MR imaging to identify and localize tumor foci at 0.5-T MR imaging.
A more objective approach is to correlate 1.5- and 0.5-T MR imaging by using deformable registration of 1.5-T MR imaging to 0.5-T MR imaging. By warping the 1.5-T MR imaging to match real-time FGRE images, a physician can have access to more detailed information about the tumor location, while retaining the real-time imaging capability of intraoperative MR imaging. This deformable registration can be later generalized for transrectal US-guided biopsy to fuse diagnostic MR imaging onto the transrectal US images.
In conclusion, T2-weighted MR images can be used to guide placement of biopsy needles in both suspected targets and sextant location in the peripheral zone of the prostate by using 3D SLICER software. While both cases indicate that the guidance with T2-weighted images is useful to ensure sextant sampling of the peripheral zone, the second case more importantly shows that the method is suitable for targeted biopsy.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, F.A.J., C.M.C.T.; study concepts, N.H., C.M.C.T., R.C., A.N., A.V.D., R.K.; study design, N.H., R.C.; literature research, N.H., S.G.S., C.M.C.T.; clinical studies, N.H., M.J., R.C., C.M.C.T., A.V.D.; experimental studies, N.H., D.K., D.G., A.N., R.K.; data acquisition, M.J., D.K., A.N., S.G.S.; data analysis/interpretation, M.J., R.C., D.K., S.G.S.; statistical analysis, N.H.; manuscript preparation and editing, N.H., C.M.C.T.; manuscript definition of intellectual content, D.K.; manuscript revision/review, N.H., M.J., D.K., R.C., D.G., S.G.S., A.V.D., R.K., F.A.J., C.M.C.T.; manuscript final version approval, F.A.J., C.M.C.T.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults DiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. C. Susil, A. Krieger, J. A. Derbyshire, A. Tanacs, L. L. Whitcomb, G. Fichtinger, and E. Atalar System for MR Image-guided Prostate Interventions: Canine Study Radiology, September 1, 2003; 228(3): 886 - 894. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
